Dear ATMs,
For an amateur working solo perhaps the biggest aperture that could be managed by him with minimum level of comfort is 30 inches. In particular during the parabolizing stage – the huge amount of glass to be removed casts a nightmare to every ATM who attempts to deal with it.
It is very unfortunate that our mirror, after polishing, turned out to be an oblate spheroid. It means that we must spent double the amount of labor to bring it back to a paraboloid. We have been working for several hours on the central part of the mirror but it seems that the progress is slow.
We have changed the testing method to Offner null test. In this arrangement the light sent out by the slit is modified by two lens to produce a negative spherical aberration. If the mirror under test is a paraboloid its positive spherical aberration should just compensates to those from theincident rays. In other words when the mirror darkens uniformly ( or with straight Ronchi bands ) the surface is already a paraboloid.
At the moment our mirror just appears to be a very exaggerate oblate, when viewed through the Offner tester. We are now being engaged into a very long battle.
Best regards
Chan Yuk Lun
8-7-2005
The making of 30 in telescope - 21, the Offner tester
-
Chanlunlun
- 夸克星
- 文章: 3847
- 註冊時間: 週四 09 10月, 2003 21:06
The making of 30 in telescope - 21, the Offner tester
- 附加檔案
-
- The light source as seen directly through the nulling lens of the Offner tester. On the left hand side of the light source is the knife edge.
- 30_34.jpg (76.11 KiB) 已瀏覽 23481 次
-
- The 30 inch polished mirror is waiting to be mounted for the Offner test. The Offner tester is on the top side of the picture.
- 30_33.jpg (51.69 KiB) 已瀏覽 23481 次
-
- The Offner tester. It has been modified so that alignment can be more easily executed during testing.
- 30_32.jpg (45.36 KiB) 已瀏覽 23481 次
Dear Chan Sir:
I'm a newbie in the field of ATM, and have tow questions about your sophisticated Offner tester:
1. Accord to my knowledge, in conventional Foucault test(without compensators), the slit and KE must strictly parallel or the sensitivity of the test would be badly downgraded. How do you manage to keep them parallel while they are some what "independently" install on to the base? Or in the case of compensate lenses are introduced, it is of different story and we just need not to care about that any more?
2. According to the photograph, the separation of the slit and knife is about several centimeters, not less than 50mm I guess. So the mirror under test must be tilt by about 0.36 degree (=arcsin(50mm/2/4000mm), in the case of the previous 24 inch project) . I did some simulations with OSLO and found that very strong coma (and astigmatism, much weaker relatively though) was introduced by the tilting, which could not be cured by any methods I knew. Could you explain why it does not matter as a guide of figuring?
2.1. As well as tilting the mirror, we can also arrange the axis of the Offner tube on the axis of the mirror thus a coaxial system was built, and place the light source off the axis instead. By this method the foci(so the KE) can also be displaced, but the coma and astigmatism are even larger than the setup described by question 2.
I'm a newbie in the field of ATM, and have tow questions about your sophisticated Offner tester:
1. Accord to my knowledge, in conventional Foucault test(without compensators), the slit and KE must strictly parallel or the sensitivity of the test would be badly downgraded. How do you manage to keep them parallel while they are some what "independently" install on to the base? Or in the case of compensate lenses are introduced, it is of different story and we just need not to care about that any more?
2. According to the photograph, the separation of the slit and knife is about several centimeters, not less than 50mm I guess. So the mirror under test must be tilt by about 0.36 degree (=arcsin(50mm/2/4000mm), in the case of the previous 24 inch project) . I did some simulations with OSLO and found that very strong coma (and astigmatism, much weaker relatively though) was introduced by the tilting, which could not be cured by any methods I knew. Could you explain why it does not matter as a guide of figuring?
2.1. As well as tilting the mirror, we can also arrange the axis of the Offner tube on the axis of the mirror thus a coaxial system was built, and place the light source off the axis instead. By this method the foci(so the KE) can also be displaced, but the coma and astigmatism are even larger than the setup described by question 2.
- 附加檔案
-
- exeggrate picture to make my questions clearer
- Offner_tester.JPG (19.45 KiB) 已瀏覽 10716 次
-
Chanlunlun
- 夸克星
- 文章: 3847
- 註冊時間: 週四 09 10月, 2003 21:06
Dear Mr/Miss bbsdma,
If you still don't understand why - just a hint to you - consider a slit is a collection a pin holes. There is no parallelism between a straight edge and a point.
Actually the Foucalt test is among one of the most beautiful experinment in Physics. The beauty of Foucalt test lies in the fact that we don't need delicate equipment to obtain yet surprisingly precise result - up to molecular level of accuracy, if such test is carried out by good hands !
To make a long story short - the light rays reflected from the mirror surface carry ALL the information concerning about the priofile of the surface under test. The knife edge actually acts as a phase-amplitude inversion filter ( well, a scary name ! ) - it filter out certain kind of information for our eye to detect and, in this case - it filters out ONLY the the information about zonal errors of the mirror surface. Other information such as coma and astigmatism are generally blocked by the filter ( the knife edge). If you still don't get it think about a nebula filter.
OSLO ( the act of ray tracing, i.e. pure geometrical optics ) does not lie to you - if you are using a pinhole to carry out the Foucalt test and employing an eyepiece to replace the knife edge all you see will be a very blurred focus with serious coma, or perhaps even some added amount of astigmatism - just the same as what has been predicted by OSLO. An eyepiece is not a filter - it let all information pass through to your eye and all these information sum up to be a blurred spot, a result by OSLO.
That's why a Foucalt test is immune to off aixs coma - because by the very act of cutting a knife edge into the light cone you are actaully performing ann information filtering process - to let through ONLY the zonal information of the mirror. Other off axis defects are not passing through. They do exist, but you don't see them against the knife edge.
Now you will perhaps see a little bit that ray tracing is far from adequate to explain some phnomena in astronomical optics. This is by no means surprising - after all we are probing so deep into the phenomena that we are actually seeing the very essence of light - its wave nature !
Best regards
Chan Yuk Lun
This is not true. Even though the slit and the knife edge are not exactly parallel, the precision of the Foucalt test is only affected minimally. Of course during operation we always try to mantain good parallelism of the knife edge and slit. One very simple method - observe the real image of the knife edge formed by the mirror and adjust the knife edge until they are parallel to your naked eye. This is more than sufficuient.bbsdma 寫: Accord to my knowledge, in conventional Foucault test(without compensators), the slit and KE must strictly parallel or the sensitivity of the test would be badly downgraded.
If you still don't understand why - just a hint to you - consider a slit is a collection a pin holes. There is no parallelism between a straight edge and a point.
Actually the Foucalt test is among one of the most beautiful experinment in Physics. The beauty of Foucalt test lies in the fact that we don't need delicate equipment to obtain yet surprisingly precise result - up to molecular level of accuracy, if such test is carried out by good hands !
Before your question is answered let me try to stress that Foucalt test cannot be fully explained by pure geometrical optic. In oder to fully understand the principle behind Foucalt test we need more advanced optics theory - the theory of Coherent Optical Information Processing. ( Fourier Optics )bbsdma 寫: According to the photograph, the separation of the slit and knife is about several centimeters, not less than 50mm I guess. So the mirror under test must be tilt by about 0.36 degree (=arcsin(50mm/2/4000mm), in the case of the previous 24 inch project) . I did some simulations with OSLO and found that very strong coma (and astigmatism, much weaker relatively though) was introduced by the tilting, which could not be cured by any methods I knew. Could you explain why it does not matter as a guide of figuring?
To make a long story short - the light rays reflected from the mirror surface carry ALL the information concerning about the priofile of the surface under test. The knife edge actually acts as a phase-amplitude inversion filter ( well, a scary name ! ) - it filter out certain kind of information for our eye to detect and, in this case - it filters out ONLY the the information about zonal errors of the mirror surface. Other information such as coma and astigmatism are generally blocked by the filter ( the knife edge). If you still don't get it think about a nebula filter.
OSLO ( the act of ray tracing, i.e. pure geometrical optics ) does not lie to you - if you are using a pinhole to carry out the Foucalt test and employing an eyepiece to replace the knife edge all you see will be a very blurred focus with serious coma, or perhaps even some added amount of astigmatism - just the same as what has been predicted by OSLO. An eyepiece is not a filter - it let all information pass through to your eye and all these information sum up to be a blurred spot, a result by OSLO.
That's why a Foucalt test is immune to off aixs coma - because by the very act of cutting a knife edge into the light cone you are actaully performing ann information filtering process - to let through ONLY the zonal information of the mirror. Other off axis defects are not passing through. They do exist, but you don't see them against the knife edge.
Now you will perhaps see a little bit that ray tracing is far from adequate to explain some phnomena in astronomical optics. This is by no means surprising - after all we are probing so deep into the phenomena that we are actually seeing the very essence of light - its wave nature !
Best regards
Chan Yuk Lun
Dear Chan Sir,
Thank you very much for your very informative answer.
I don't think I have the ability to ask any further questions about test theory soon, since it'll take me several months or even several years to learn enough about Fourier Optics to understand your teaching . Thus I still have some minor pratical ones:
1. You said the KE filters out other errors but zonal ones. So we can not use the apparatus alone to detect the asymmetric errors of the mirrors itself?
Can we just turn the mirror by 90 degrees and read the shadow again then compare the difference of the two zonal shadows(before and after the mirror rotation), thus we can tell the mirror is symmetric(if they are the same) or not(if they are not the same)?
2.The width of the slit you chose, and how does it affect the test sensibility?
3. Some advanced mirror makers in mainland China are very interested of your Offner tester, but questions similar to mine arises. So may I translate your answer above into Chinese and let your ideas known among the ATM community of China?
Thank you very much for your very informative answer.
I don't think I have the ability to ask any further questions about test theory soon, since it'll take me several months or even several years to learn enough about Fourier Optics to understand your teaching . Thus I still have some minor pratical ones:
1. You said the KE filters out other errors but zonal ones. So we can not use the apparatus alone to detect the asymmetric errors of the mirrors itself?
Can we just turn the mirror by 90 degrees and read the shadow again then compare the difference of the two zonal shadows(before and after the mirror rotation), thus we can tell the mirror is symmetric(if they are the same) or not(if they are not the same)?
2.The width of the slit you chose, and how does it affect the test sensibility?
3. Some advanced mirror makers in mainland China are very interested of your Offner tester, but questions similar to mine arises. So may I translate your answer above into Chinese and let your ideas known among the ATM community of China?
to蘇魯錠先生：
小弟未懂光學測試之萬一，鬥膽為兩位前輩的溝通出點棉力。 您所提的問題， 事實上小弟在樓上幾個帖子剛剛向陳先生請教過，粗略繙譯如下：
bbsdma 向陳先生請教：
從照片看，縫和刀口距離數公分，估計不下50mm。這樣被測主鏡至少要傾斜0.36度（陳先生之前的24寸主鏡的情況）。我在OSLO中做了一些模擬，這個傾斜會導緻非常強的慧差（和相對較弱的像散），這種像差不能通過我所知道的任何方法處理。您能解釋為何（這）對我們的檢測沒有影響呢？
陳先生答復：
在我開始回答前，我必須強調 佛科測試不能用幾何光學作完全的解釋。要完整理解佛科測試的原理，我們需要更高深的光學理論：相幹光信息處理（傅立葉光學）
長話短說，從被測主鏡反射回來的光綫擕帶着鏡面的【所有】信息。刀口實際上扮縯了 相位－強度轉換濾波器。它為我們的眼睛過濾出一些特定種類的信息，在（我們正在討論的這個測試儀的）情況下，隻有鏡面環帶錯誤的信息能通過。其他的信息，比如慧差和像散，則基本被濾波器（刀口）擋住了。 如果你仍然不明白，可以想一下星雲濾鏡。
OSLO（光綫追蹤，純粹幾何光學）不會騙你。如果你用針孔進行彿科測試，然后用目鏡而不是刀口觀察，你將隻會見到非常糢糊的、帶有強烈慧差和像散的焦點，一如OSLO所預測。 目鏡不是濾波器， 它讓所有的信息通過，到達你的眼睛：所有這些信息曡加起來，就是一個糢糊的點，即是OSLO的結果。
這是為什么佛科測試對離軸慧差免疫的原因：在用刀口切入光錐的時候，你實際進行着光信息過濾過程——只允許鏡面的環帶誤差信息通過。其他離軸缺陷不能通過。它們確實存在，但你不能在刀口后面看到它們。
現在你或許領悟到，光綫追蹤遠遠未能解析天文光學上的所有現象。這毫不奇怪，畢竟我們在深入地探討這個現象——實際上我們看到光的本質——它的波動天性！（譯註：這句話我看不太懂，姑且先看）
小弟未懂光學測試之萬一，鬥膽為兩位前輩的溝通出點棉力。 您所提的問題， 事實上小弟在樓上幾個帖子剛剛向陳先生請教過，粗略繙譯如下：
bbsdma 向陳先生請教：
從照片看，縫和刀口距離數公分，估計不下50mm。這樣被測主鏡至少要傾斜0.36度（陳先生之前的24寸主鏡的情況）。我在OSLO中做了一些模擬，這個傾斜會導緻非常強的慧差（和相對較弱的像散），這種像差不能通過我所知道的任何方法處理。您能解釋為何（這）對我們的檢測沒有影響呢？
陳先生答復：
在我開始回答前，我必須強調 佛科測試不能用幾何光學作完全的解釋。要完整理解佛科測試的原理，我們需要更高深的光學理論：相幹光信息處理（傅立葉光學）
長話短說，從被測主鏡反射回來的光綫擕帶着鏡面的【所有】信息。刀口實際上扮縯了 相位－強度轉換濾波器。它為我們的眼睛過濾出一些特定種類的信息，在（我們正在討論的這個測試儀的）情況下，隻有鏡面環帶錯誤的信息能通過。其他的信息，比如慧差和像散，則基本被濾波器（刀口）擋住了。 如果你仍然不明白，可以想一下星雲濾鏡。
OSLO（光綫追蹤，純粹幾何光學）不會騙你。如果你用針孔進行彿科測試，然后用目鏡而不是刀口觀察，你將隻會見到非常糢糊的、帶有強烈慧差和像散的焦點，一如OSLO所預測。 目鏡不是濾波器， 它讓所有的信息通過，到達你的眼睛：所有這些信息曡加起來，就是一個糢糊的點，即是OSLO的結果。
這是為什么佛科測試對離軸慧差免疫的原因：在用刀口切入光錐的時候，你實際進行着光信息過濾過程——只允許鏡面的環帶誤差信息通過。其他離軸缺陷不能通過。它們確實存在，但你不能在刀口后面看到它們。
現在你或許領悟到，光綫追蹤遠遠未能解析天文光學上的所有現象。這毫不奇怪，畢竟我們在深入地探討這個現象——實際上我們看到光的本質——它的波動天性！（譯註：這句話我看不太懂，姑且先看）
-
Chanlunlun
- 夸克星
- 文章: 3847
- 註冊時間: 週四 09 10月, 2003 21:06
Dear Mr. So,苏鲁锭 寫:业余做offner补偿检测最大的困难就是补偿透镜的折射率很难测到规定的要求。
当补偿透镜的折射率在1.5左右时，折射率误差0.0000001，检测的误差就是1/100波长。
一般很难测到0.000001.那么检测误差就有1/10波长了。
业余检测可能用反射补偿更好一点。
You are telling the truth. High quality precision lens is not readily offered to amateur hands.
Offner test only needed when dealing with mirror of very big size ( 24 inches or more ) and very fast f/ratio. ( f/3.0 pr faster ). Otherwise a classical Foucalt zonal test is good enough in most situations.
Actually with the help of a computer the pecision reached by Foucalt test is already very high, so high that it practically tells you zones of less than 1/10 wave , good enogh to figure mirrors of alomst all configuration.
Best regards
Chan Yuk Lun
-
Chanlunlun
- 夸克星
- 文章: 3847
- 註冊時間: 週四 09 10月, 2003 21:06
Dear Mr. So,苏鲁锭 寫:您讲的有道理。
但是刀口仪是可以看到彗差的。
我正在用刀口仪检测抛物面，一般口径203焦距800时刀口仪位置偏差了0.2MM是就可以看到明显的彗差。
May be that you have a pair of superior eyes.
I have figured more than several hundred pieces of mirror and have not observed coma of any kinds in a Foucalt test.
Recently I have re-examined our 24 inch f/3.5 and my 20 inches f/4 by Foucalt test and again coma was absent in my observation.
By the very words "Foucalt test" I mean using a knife edge, not an eyepiece. If you were using an eyepiece in the center of curvature you will then observe serious coma. But when you are using a knife edge situation are not the same.
Best regards
Chan Yuk Lun
[quote="Chanlunlun
Dear Mr. So,
You are telling the truth. High quality precision lens is not readily offered to amateur hands.
Offner test only needed when dealing with mirror of very big size ( 24 inches or more ) and very fast f/ratio. ( f/3.0 pr faster ). Otherwise a classical Foucalt zonal test is good enough in most situations.
Actually with the help of a computer the pecision reached by Foucalt test is already very high, so high that it practically tells you zones of less than 1/10 wave , good enogh to figure mirrors of alomst all configuration.
Best regards
Chan Yuk Lun[/quote]
粗淺的翻譯一下：
陳先生致蘇先生：
你所言不虛。對業余人士來說，并無現成的高質量的精密透鏡供應。
只有當需要應付非常大（24英寸或以上）和非常快（F/3.0或更快）的主鏡的時候，才需要進行Offner測試。其他大多數情況下，一個經典的傅科區域測試法已經足夠好了。
事實上，在電腦的協助下，傅科測試所達致的精度是非常高的，實踐上能夠指出小于1/10波長的環帶誤差， 對修正多數類型的鏡面來說，這樣的精度已經是足夠的了。
致意！
Dear Mr. So,
You are telling the truth. High quality precision lens is not readily offered to amateur hands.
Offner test only needed when dealing with mirror of very big size ( 24 inches or more ) and very fast f/ratio. ( f/3.0 pr faster ). Otherwise a classical Foucalt zonal test is good enough in most situations.
Actually with the help of a computer the pecision reached by Foucalt test is already very high, so high that it practically tells you zones of less than 1/10 wave , good enogh to figure mirrors of alomst all configuration.
Best regards
Chan Yuk Lun[/quote]
粗淺的翻譯一下：
陳先生致蘇先生：
你所言不虛。對業余人士來說，并無現成的高質量的精密透鏡供應。
只有當需要應付非常大（24英寸或以上）和非常快（F/3.0或更快）的主鏡的時候，才需要進行Offner測試。其他大多數情況下，一個經典的傅科區域測試法已經足夠好了。
事實上，在電腦的協助下，傅科測試所達致的精度是非常高的，實踐上能夠指出小于1/10波長的環帶誤差， 對修正多數類型的鏡面來說，這樣的精度已經是足夠的了。
致意！
[quote="Chanlunlun"][quote="苏鲁锭"]您讲的有道理。
但是刀口仪是可以看到彗差的。
我正在用刀口仪检测抛物面，一般口径203焦距800时刀口仪位置偏差了0.2MM是就可以看到明显的彗差。[/quote]
Dear Mr. So,
May be that you have a pair of superior eyes.
I have figured more than several hundred pieces of mirror and have not observed coma of any kinds in a Foucalt test.
Recently I have re-examined our 24 inch f/3.5 and my 20 inches f/4 by Foucalt test and again coma was absent in my observation.
By the very words "Foucalt test" I mean using a knife edge, not an eyepiece. If you were using an eyepiece in the center of curvature you will then observe serious coma. But when you are using a knife edge situation are not the same.
Best regards
Chan Yuk Lun[/quote]
粗淺的翻譯如下：
陳先生致蘇先生：
可能你有一對著異乎尋常的好眼睛。
我曾經修改過數百塊鏡面，未有任何形式的慧差被觀察到。
最近我用傅科測試法重新檢驗了我們的24英寸F/3.5和20英寸F/4，再次地，慧差沒有被我觀察到。
我所指的“傅科測試法”是用刀口而不是目鏡。如果你在曲率中心用目鏡觀察，你將會看到嚴重的慧差。但當你用刀口， 情況則不一樣。
致意！
但是刀口仪是可以看到彗差的。
我正在用刀口仪检测抛物面，一般口径203焦距800时刀口仪位置偏差了0.2MM是就可以看到明显的彗差。[/quote]
Dear Mr. So,
May be that you have a pair of superior eyes.
I have figured more than several hundred pieces of mirror and have not observed coma of any kinds in a Foucalt test.
Recently I have re-examined our 24 inch f/3.5 and my 20 inches f/4 by Foucalt test and again coma was absent in my observation.
By the very words "Foucalt test" I mean using a knife edge, not an eyepiece. If you were using an eyepiece in the center of curvature you will then observe serious coma. But when you are using a knife edge situation are not the same.
Best regards
Chan Yuk Lun[/quote]
粗淺的翻譯如下：
陳先生致蘇先生：
可能你有一對著異乎尋常的好眼睛。
我曾經修改過數百塊鏡面，未有任何形式的慧差被觀察到。
最近我用傅科測試法重新檢驗了我們的24英寸F/3.5和20英寸F/4，再次地，慧差沒有被我觀察到。
我所指的“傅科測試法”是用刀口而不是目鏡。如果你在曲率中心用目鏡觀察，你將會看到嚴重的慧差。但當你用刀口， 情況則不一樣。
致意！
誰在線上
正在瀏覽這個版面的使用者：沒有註冊會員 和 12 位訪客